Model Development and Validation of a Dual-Axis PV Tracking System: A Case of South Africa

 
CONTINUE READING
Model Development and Validation of a Dual-Axis PV Tracking System: A Case of South Africa
International Journal of Electrical and Electronic Engineering & Telecommunications Vol. 10, No. 4, July 2021

          Model Development and Validation of a
         Dual-Axis PV Tracking System: A Case of
                      South Africa
                       Percy Andrew Hohne1, Kanzumba Kusakana1, and Bubele Papy Numbi2
 1
     Dept. of Electrical, Electronic and Computer Engineering, Central Univ. of Technology, Bloemfontein, South Africa
            2
              Dept. of Electrical Engineering, Mangosuthu University of Technology, Durban 4031, South Africa
                             Email: phohne@cut.ac.za; kkusakana@cut.ac.za; numbib@mut.ac.za

Abstract—Dual axis photovoltaic (PV) tracking system is                     In hindsight, the rise in electricity prices along with
considered in general to be a poor investment. This is                   time-based pricing and maximum demand penalties has
mainly due to the substantial initial investment costs that              resulted in exceedingly high grid energy costs. The rise in
these systems carry. However, in recent years, solar panels              grid energy costs combined with reduced implementation
and accompanying component costs have decreased
significantly. Additionally, electricity price hikes in South
                                                                         costs of renewable energy systems has improved the
Africa have compelled most of the country’s citizens to                  feasibility of these systems [4].
reconsider their sources of electrical energy. A popular                    Solar energy systems in particular, have been a proven
alternative to grid energy in South Africa is the use of                 and widely used method of alternative energy generation
photovoltaic systems. Careful consideration is required                  in most areas in South Africa, as opposed to wind energy
when choosing from the various systems available on the                  harnessing technologies [5]. Various methods and
market. The main method for maximizing the output power                  technologies have emerged to increase the photovoltaic
of these systems is to introduce solar tracking systems.
Therefore, in this paper, a model of a dual axis tracking                energy yield of photovoltaic (PV) modules. These include;
system is developed and validated against a real-world plant             enhancements in inverter technologies for maximum
in the Bloemfontein region in South Africa. The presented                power point tracking (MPPT), cooling of PV modules to
model was observed to be accurate to within an error rate of             increase the efficiency of the modules and solar tracking
6.39%. Additionally, the performance of the inverters of the             systems to absorb maximum energy from the solar
PV tracking systems were evaluated and discussed. The                    resource [6]. Substantial research has been conducted on
validated model may prove to be an excellent tool for energy
managers to determine the feasibility of such systems,
                                                                         improving the MPPT capability of inverters and cooling
compared to conventional photovoltaic setups.                            PV modules for increased efficiency as in [7], [8], which
                                                                         resulted in significant advances for PV power production.
Index Terms—Dual axis PV tracking system, model
                                                                         Similarly, solar tracking systems have been subjected to
development, model validation, PV performance analysis
                                                                         extensive experimentation. In most cases, a single axis
                                                                         tracking system offered sufficient performance given the
                        I. INTRODUCTION                                  price bracket for installing such a system [9]. Dual axis
                                                                         system, on the other hand, is often too costly in terms of
   In South Africa, the prevalence of renewable energy                   the initial investment required for implementation.
systems for power generation has increased tremendously                  However, in recent years, with the decline in PV module
in the last decade [1]. This is mainly due to the significant            costs and rise in electricity prices, these systems have
rise in electricity prices and the electricity supplier’s                become increasingly competitive with conventional
inability to meet the energy demand of consumers [2]. As                 stationary PV and single axis PV tracking systems.
a result, the electricity supplier, Eskom, has introduced                Additionally, it may be necessary in some cases to
load shedding or curtailment in order to mitigate a total                implement Dual axis tracking systems where the physical
grid shutdown or blackout. Additional strategies to
                                                                         installation space is limited. Other benefits of these
reduce the strain on the national grid are the implementa-
                                                                         systems include PV power production during the costly
tion of time-based energy pricing and maximum demand
                                                                         regions of time-based pricing schemes [10].
penalties [3]. Time based energy pricing serves as an
                                                                            In retrospect, it may seem difficult to compare these
incentive to consumers to exercise demand side
                                                                         systems without accurate models to predict their
management, while maximum demand penalties are
                                                                         performance under certain weather conditions in various
enforced when an upper demand limit is reached. 
                                                                         regions. Several studies have been conducted on the
                                                                         model development of small-scale PV tracking systems
                                                                         as in [11]-[13]. However, for the specific case of South
   Manuscript received September 10, 2020; revised November 5, 2020;
accepted December 18, 2020.
                                                                         Africa, few studies have focused on large scale PV
   Corresponding author: P. A. Hohne (email: phohne@cut.ac.za).          tracking system performance, particularly in the central

©2021 Int. J. Elec. & Elecn. Eng. & Telcomm.                       288
doi: 10.18178/ijeetc.10.4.288-293
Model Development and Validation of a Dual-Axis PV Tracking System: A Case of South Africa
International Journal of Electrical and Electronic Engineering & Telecommunications Vol. 10, No. 4, July 2021

region of the country. Therefore, in this paper, a model is                    units combined may deliver a total power of 151.2kWp.
developed and validated against the operation of a large-                      Each of the tracking units are fitted with a 15KVA
scale solar tracking plant, consisting of multiple                             inverter with MPPT capability. Five of these tracking
standalone tracking units, in the Bloemfontein region in                       units are shown in Fig. 1. In Fig. 2, the weather station,
South Africa. The aim is to provide an accurate model to                       located near the PV tracking plant is shown. The data
forecast the PV power production of these systems in                           from the weather station may be used for model
order to compare the feasibility as opposed to                                 validation purposed, more details of the station are
conventional PV systems.                                                       provided further in this section.
                                                                                  On average, the tracking units yield approximately the
                   II. MODEL DEVELOPMENT                                       same amount of energy throughout the day with slight
                                                                               variances (
Model Development and Validation of a Dual-Axis PV Tracking System: A Case of South Africa
International Journal of Electrical and Electronic Engineering & Telecommunications Vol. 10, No. 4, July 2021

                                         TABLE I: MONTHLY PV OUTPUT POWER YIELDS FOR 2019 [MWH]
Inverter      Jan.      Feb.     Mar.       Apr.     May     June          July    Aug.     Sep.    Oct.     Nov.    Dec.    Average
     01       4,06      5,67     3,38       1,03     4,96     2,99         5,21     5,82    1,16    2,09     7,4     5,21     4,08
     02       0,24      3,45     2,62       1,39     0,08     0,61         3,28     0,15    0,38    0,55     3,48    3,28     1,63
     03       3,61      2,23     0,16       2,18     2,87     4,09         0,27     4,74    4,48    1,6      1,14    0,27      2,3
     04       2,37      3,44     4,24       2,33     1,5      2,48         2,13     1,41    0,15    2,58     1,86    2,13     2,22
     05       2,15      2,26     2,14       0,59     0,28     2,15         2,15     1,61    1,38    0,15     2,39    2,15     1,62
     06       2,83      0,23     0,87       0,94     3,08     3,52         0,92     5,31    2,07    0,96     1,2     0,92      1,9
     07       8,91       0       1,73       2,14     7,15    11,32         1,34    10,71    0,39    1,18     1,85    1,34     4,01
     08       0,91      3,35     1,82       0,51     1,27     0,14         2,84     3,13    1,84    0,68     3,76    2,84     1,92
     09       2,92      5,36     4,25       2,95     2,65     1,49         2,34     5,15    3,46    4,3      4,04    2,34     3,44
     10       4,23      1,19      3         3,24     1,75     4,85         2,85     2,97    1,65    2,74     1,53    2,85     2,74
     11       4,84      3,58     2,15       0,33     1,48     3,54         1,69     3,75     0      3,97     1,69    1,69     2,39
     12       2,65      1,4      4,78       1,5      0,88     1,58         3,68     0,81    0,29    2,55     2,11    3,68     2,16

                                          TABLE II: PV OUTPUT POWER DEVIATION FROM AVERAGE [%]
Inverter       Jan.     Feb.     Mar.       Apr.    May      June       July       Aug.     Sep.     Oct.   Nov.     Dec.    Average
      1       2,518     2,62     2,806       2,3    2,613    2,503     2,482       2,851   2,876    3,209   3,015    2,482    2,69
      2       2,421    2,859     2,869      2,256   2,486    2,381     2,845       2,768   2,919    3,463   3,433    2,845     2,8
      3       2,521     2,75     2,867       2,3    2,573    2,487     2,729       2,875   2,968    3,386   3,291    2,729    2,79
      4       2,487    2,855     2,914      2,277   2,531    2,461     2,815       2,102   2,926    3,448   3,363    2,815    2,75
      5       2,489    2,831     2,842      2,235   2,509    2,439     2,807       2,766   2,873    3,442    3,39    2,807    2,79
      6       2,494    2,795     2,852      2,286   2,572    2,463     2,768       2,856   2,918    3,431   3,351    2,768     2,8
      7       2,203    2,815     2,877      2,204    2,23    2,165     2,795       2,623   2,939    3,446   3,377    2,795    2,71
      8       2,411     2,84     2,837      2,217   2,476    2,381     2,836       2,729   2,895     3,46   3,418    2,836    2,78
      9       2,372     2,66     2,698      2,169   2,432    2,333     2,692       2,665   2,782    3,265   3,188    2,692    2,66
     10       2,307    2,699     2,728      2,179   2,412    2,308     2,677       2,645   2,816    3,299   3,251    2,677    2,67
     11       2,511    2,798     2,854      2,232    2,56    2,507     2,809       2,738   2,927    3,461   3,367    2,809     2,8
     12       2,329    2,654     2,729      2,203   2,464    2,377     2,587       2,661   2,853    3,254   3,204    2,587    2,66

   It may be argued that the PV array orientation is                       inverter 9 shows a high deviation from the average, while
dependent both on the manufacturer and environmental                       inverter 11 shows a comparatively low deviation.
conditions. The solar tracking sensor and tracking                            In hindsight, the deviation shows little correlation with
actuators may have certain tolerances of operation from                    the PV tracking array’s performance. However, the
the manufacturer in which the orientation of each array                    deviation may be used to determine the accuracy of the
can vary with respect to each other. In addition,                          model given the average power produced from all the
considering environmental conditions, accumulation of                      tracking arrays.
dust and dirt on the sensors and actuators along with                         Therefore, for model validation purposes, two tracking
scattered cloud conditions may similarly affect the                        systems are identified for each month that produced the
performance of each tracking array.                                        highest output power. This would minimize the
   In Table I, the monthly PV output power yields are                      uncertainties in the model, such as all the factors that
shown. The PV arrays with the highest output power is                      affect PV output performance as discussed previously in
highlighted in green in groups of two, while the lowest                    this section.
output power for each array is highlighted in red. From                       The power yield readings from the energy monitoring
the table, it may be observed that inverters 1 to 6 had the                dashboard of these existing (real world) systems, logged
highest concentration in maximum power output                              at 5-minute intervals were used as historical data to
throughout the year, while inverters 7 to 12 showed poor                   validate the PV tracking model.
performance in comparison.                                                    Exogenous data obtained from a weather station
   In hindsight, the highest annual average in power                       located approximately 250 m from the solar tracking
production was observed to be Inverter 6 and Inverter 11,                  arrays at the Central University of Technology (CUT)
while Inverter 9 and Inverter 12 had the lowest power                      were fed into the developed PV tracking model to
yield for the year.                                                        determine the output power for comparison with the
   In Table II, the deviation in PV output power from the                  existing system [17]. The weather station at the CUT
overall average power produced is shown on a monthly                       forming part of the South African Universities
basis. In this Table, the lowest deviation is highlighted in               Radiometric Network (SAURAN) is shown in Fig. 2.
green, while the highest is shown in red.                                     The data acquired from the weather station are noted in
   From the table, it may be observed that Inverter 1 and                  (1) to (3). These include; solar irradiance (global
Inverter 7, deviated substantially from the average with                   horizontal, direct normal, diffuse horizontal irradiance),
Inverter 2 and Inverter 5 showing the lowest deviation.                    ambient air temperature, azimuth angle, tilt angle of the
   Comparing the maximum power output with the                             irradiance sensor.
deviation reveals that the inverters that produced the                        In order to simulate the operation of the dual axis
highest power have relatively low deviation from the                       tracking PV array, the tilt angle noted by (β) in Eq. (3)
average power produced. However, on the lower                              was equated to the calculated tilt angle of the solar
performing units, a mixed result is observed, where                        irradiance measurement system of the weather station.

©2021 Int. J. Elec. & Elecn. Eng. & Telcomm.                         290
International Journal of Electrical and Electronic Engineering & Telecommunications Vol. 10, No. 4, July 2021

   Fig. 3. Actual tracking system vs. model for a selected day in Jan.           Fig. 9. Actual tracking system vs. model for a selected day in Jul.

    Fig. 4. Actual tracking system vs. model for a selected day in Feb.         Fig. 10. Actual tracking system vs. model for a selected day in Aug.

    Fig. 5. Actual tracking system vs. model for a selected day in Mar.         Fig. 11. Actual tracking system vs. model for a selected day in Sept.

    Fig. 6. Actual tracking system vs. model for a selected day in Apr.         Fig. 12. Actual tracking system vs. model for a selected day in Oct.

   Fig. 7. Actual tracking system vs. model for a selected day in May.          Fig. 13. Actual tracking system vs. model for a selected day in Nov.

    Fig. 8. Actual tracking system vs. model for a selected day in Jun.         Fig. 14. Actual tracking system vs. model for a selected day in Dec.

©2021 Int. J. Elec. & Elecn. Eng. & Telcomm.                              291
International Journal of Electrical and Electronic Engineering & Telecommunications Vol. 10, No. 4, July 2021

   In Fig. 3 to Fig. 14, the comparisons between the                                           CONFLICT OF INTEREST
simulated operation (model) and the actual data (real-
                                                                              The authors declare no conflict of interest.
world) of the system are illustrated for each month. This
method follows historical data validation procedures to
ensure accurate real-world representation of simulated                                       AUTHOR CONTRIBUTIONS
results.                                                                   Percy A. Hohne conducted the research and analyzed
   From Fig. 3, Fig. 4, Fig. 13 and Fig. 14, it may be                  the data; Percy A. Hohne and Kanzumba Kusakana wrote
observed that the PV output power trend varies                          the paper; Bubele P. Numbi assisted in the model
substantially as oposed to the output power illustrated in              development; all authors approved the final version.
Fig. 5 to Fig. 12. This is mainly attributed to the
overcast/cloudy skies, frequently experienced during the                                           ACKNOWLEDGEMENT
summer months in South Africa. In hindsight, the error in
PV output power increases for these scenarios where                        The authors would like to thank the Central University
overcast skies are apparent. This increase in error is                  of Technology for data pertaining to the study and
unavoidable as each tracking system receives different                  financial support.
amounts of solar irradiance due to scattered cloud cover.
   In retrospect, comparisons between the simulated                                                      REFERENCES
results and the actual data of the PV tracking systems                  [1]    S. O. Oyedepo, “Energy and sustainable development in Nigeria:
reveal that the model represents real-world operation of                       The way forward,” Energy, Sustainability and Society, vol. 2, July
the system within a margin of error. In Table III, the                         2012.
observed error for each month is depicted.                              [2]    P. A. Hohne, K. Kusakana, and B. P. Numbi, “Improving energy
   The average error over the evaluated 12-month period                        efficiency of thermal processes in healthcare institutions: A review
                                                                               on the latest sustainable energy management strategies,” Energies,
was observed to be 7.56%. The average deviation of the                         vol. 13, no. 3, p. 569, Jan. 2020.
inverter outputs compared to the average output power of
                                                                        [3]    P. A. Hohne, K. Kusakana, and B. P. Numbi, “A review of water
the entire solar tracking plant was observed to be 2.53%                       heating technologies: An application to the South African
from Table II. Combining the error rate of the model with                      context,” Energy Reports, vol. 5, pp. 1–19, Nov. 2019.
the deviation rate of the inverters, to provide a worst case            [4]    B. E. Türkay and A. Y. Telli, “Economic analysis of standalone
scenario deviation, may result in an overall error rate of                     and grid connected hybrid energy systems,” Renewable Energy,
                                                                               vol. 36, no. 7, pp. 1931–1943, Jul. 2011.
< 9%. This may be considered as a large percentage,
however, given the numerous factors that influence the                  [5]    A. K. Aliyu, B. Modu, and C. W. Tan, “A review of renewable
                                                                               energy development in Africa: A focus in South Africa, Egypt and
PV output power of the tracking systems, these figures                         Nigeria,” Renewable and Sustainable Energy Reviews, vol. 81, pp.
may fall within an acceptable range to predict the                             2502–2518, Jan. 2018.
performance of a dual axis tracking system. The model                   [6]    B. Parida, S. Iniyan, and R. Goic, “A review of solar photovoltaic
may therofore be recommended for accurate performance                          technologies,” Renewable and Sustainable Energy Reviews, vol.
and economic predictions of PV tracking systems in the                         15, no. 3, pp. 1625–1636, Apr. 2011.
Free-state region in South Africa and by extention the rest             [7]    A. Mohapatra, B. Nayak, P. Das, and K. B. Mohanty, “A review
                                                                               on MPPT techniques of PV system under partial shading
of the world.                                                                  condition,” Renewable and Sustainable Energy Reviews, vol. 80,
                                                                               pp. 854–867, Dec. 2017.
             TABLE III: PV OUTPUT POWER ERROR [%]
                                                                        [8]    C. I. Ferreira and D. S. Kim, “Techno-economic review of solar
    Month       Jan.     Feb.     Mar.       Apr.   May    Jun.                cooling technologies based on location-specific data,”
   % Error      9.7      8.2       4.7       7.8    3.8     5.2                International Journal of Refrigeration, vol. 39, pp. 23–37, Mar.
    Month       Jul.     Aug.     Sept.      Oct.   Nov.   Dec.                2014.
   % Error      3.6      3.7       4.8       7.9    9.1     8.2         [9]     N. A. Kelly and T. L. Gibson, “Improved photovoltaic energy
   Average                            6.39                                     output for cloudy conditions with a solar tracking system,” Solar
                                                                               Energy, vol. 83, no. 11, pp. 2092–2102, Nov. 2009.
                                                                        [10]   N. R. Darghouth, R. H. Wiser, and G. Barbose, “Customer
                       IV. CONCLUSION                                          economics of residential photovoltaic systems: Sensitivities to
                                                                               changes in wholesale market design and rate structures,”
   A mathematical model of a dual axis PV tracking                             Renewable and Sustainable Energy Reviews, vol. 54, pp. 1459–
system has been developed and validated against real                           1469, Feb. 2016.
world data. The real-world data was obtained from 12                    [11]   M. Lehloka, J. Swart, and P. Hertzog, “A comparison of two
identical dual axis tracking systems located in the                            automatic solar tracking algorithms,” in Proc. E3S Web of
                                                                               Conferences, 2020, vol. 152, p. 02009.
Bloemfontein area in South Africa. The results showed
                                                                        [12]   H. J. Vermaak, “Techno-economic analysis of solar tracking
that the model accurately represented the output power                         systems in South Africa,” Energy Procedia, vol. 61, pp. 2435–
curve of an existing system within an error range of                           2438, 2014.
6.39 %. In addition, the performance of the identical                   [13]   T. Huld, M. Šúri, and E. D. Dunlop, “Comparison of potential
tracking systems was evaluated with respect to the                             solar electricity output from fixed-inclined and two-axis tracking
monthly inverter power outputs of each tracking system.                        photovoltaic modules in Europe,” Progress in Photovoltaics:
                                                                               Research and Applications, vol. 16, no. 1, pp. 47–59, Sep. 2007.
The highest deviation from the average power output of
                                                                        [14]   B. P. Numbi and S. J. Malinga, “Optimal energy cost and
the tracking units were 4.08%, while the lowest deviation                      economic analysis of a residential grid-interactive solar PV
was noted to be 1.62%. An average deviation of 2.53%                           system- case of eThekwini municipality in South Africa,” Applied
was observed.                                                                  Energy, vol. 186, pp. 28–45, Jan. 2017.

©2021 Int. J. Elec. & Elecn. Eng. & Telcomm.                      292
International Journal of Electrical and Electronic Engineering & Telecommunications Vol. 10, No. 4, July 2021

[15] Y. Riffonneau, S. Bacha, F. Barruel, and S. Ploix, “Optimal power                                   Kanzumba Kusakana is a professor in
     flow management for grid connected PV systems with batteries,”                                      electrical engineering and the Head of the
     IEEE Trans. on Sustainable Energy, vol. 2, no. 3, pp. 309–320, Jul.                                 Electrical,   Electronic     and     Computer
     2011.                                                                                               Engineering Department at the Central
[16] P. A. Hohne, K. Kusakana, and B. P. Numbi, “Optimal energy                                          University of Technology, Free State. He is
     management and economic analysis of a grid-connected hybrid                                         well-received as a professional development
     solar water heating system: A case of Bloemfontein, South                                           short course instructor in South Africa based
     Africa,” Sustainable Energy Technologies and Assessments, vol.                                      on his technical, consulting experience and
     31, pp. 273–291, Feb. 2019.                                                                         academic qualifications. With over 150
                                                                                                         publications     in    journal,     conference
[17] M. J. Brooks, et al., “SAURAN: A new resource for solar
                                                                               proceedings and book chapters, his current research looks at small scale
     radiometric data in Southern Africa,” Journal of Energy in
                                                                               renewable power generation as well as optimal energy management;
     Southern Africa, vol. 26, no. 1, pp. 2–10, Apr. 2017.
                                                                               which support the UNESCO Sustainable Development Goals. He is
[18] P. G. McCormick and H. Suehrcke, “The effect of intermittent              currently an associate editor for the IET Renewable Power Generation
     solar radiation on the performance of PV systems,” Solar Energy,          Journal and has an H-index of 25. He is a senior member of the South
     vol. 171, pp. 667–674, Sep. 2018.                                         African Institute of Electrical Engineers (SAIEE); a Professional
                                                                               Engineer (Pr Eng) registered with ECSA a NRF rated researcher.
Copyright © 2021 by the authors. This is an open access article                He received his doctorate degree in electrical engineering in 2015 from
distributed under the Creative Commons Attribution License (CC BY-             the Central University of Technology, Free State, South Africa. He also
NC-ND 4.0), which permits use, distribution and reproduction in any            holds a degree in electromechanical engineering from the University of
medium, provided that the article is properly cited, the use is non-           Lubumbashi (2006); a degree in electrical power engineering from
commercial and no modifications or adaptations are made.                       Tshwane University of Technology (2009), as well as an MBA in
                                                                               Energy Management (2019) from the Cyprus Institute of Marketing
                         Percy A. Hohne obtained his B.Tech. degree            (BVI).
                         in Electrical Engineering in 2016 and his
                         M.Eng. degree in 2018 from the Central                                         Bubele P. Numbi obtained his Ph.D degree in
                         University of Technology. Since 2019 he has                                    Elelctrial Engineeinrg, from the University of
                         been working at the Central University of                                      Pretoria (UP) in 2015. He is a NRF rated
                         Technology as a lecturer and enrolled for the                                  research. His research areas are power system
                         D.Eng. degree in Electrical Engineering. His                                   dynamics and control, energy management,
                         area of interest is optimal energy management                                  energy efficiency, and renewable energy. Dr
                         of hybrid renewable energy systems.                                            Numbi is currently working as a senior
                                                                                                        lecturer in the Department of Electrical
                                                                                                        Engineering at the MUT.

©2021 Int. J. Elec. & Elecn. Eng. & Telcomm.                             293
You can also read